Skip to content
Deliberate Signals

A Beginner’s Guide to Pendo Data Segmentation

A Beginner's Guide to Pendo Data Segmentation

If you just opened Pendo and saw the word segment for the first time, you might feel like you need a data background before you can click anything. You do not. Pendo data segmentation is a way of asking a specific question about a specific group of people instead of staring at everyone at once. Once you see it that way, the technical feeling goes away fast.

Most people who inherit Pendo or set it up fresh get stuck here because nobody showed them the order of operations. They try to build a perfect segment on day one, covering every condition they can think of. That is backwards. Start small, get one useful answer, then build from there.

What a segment actually is

A segment in Pendo is a saved filter: a group of visitors or accounts defined by rules. Pendo’s segments article lists the rule types: metadata (who someone is), product usage (what they did on which pages and features), event properties, and a CSV upload of IDs. The same segment filters analytics and targets guides, which is the whole reason it is worth building one carefully.

Think of a filing cabinet at a law office. You do not dump every document into one drawer and dig through it each time. You label drawers by case, by client type, by status. A segment is the label. The data underneath does not change. You are choosing which drawer to open.

Can you name one group of users you ask about more than once a month? If yes, that is your first segment.

Pick the data points that matter

Before you build anything, decide what you need to know. Most teams need a handful of fields to start, not forty. The ones that matter for almost every B2B product:

  • Account plan or tier, so you are not mixing free users with your highest-paying accounts.
  • Role, because an admin behaves differently than an end user.
  • Signup or activation date, because new users and veterans do not need the same view.
  • Last active date, which tells you who is still using the product.

Pick three, maybe four. You can add fields later, but you cannot get useful answers out of a segment built on data nobody defined first. Those fields arrive as metadata, the information your install snippet or a CRM sync attaches to each visitor and account. If role is not being passed in yet, the next section is how you get it anyway.

A real Pendo data segmentation example: role, without engineering

A records-retrieval product used by law firms, owned by a healthcare data company I work with, had no role field in its visitor metadata. The engineering queue was months long. So we asked users instead: a one-question in-app guide on first visit, with three choices, paralegal, legal assistant, or attorney, plus other.

47.7% of visitors answered, 1,398 of 2,933 in the first stretch, and about 90% identified as legal-side staff. The answers were saved as metadata and became the first three segments the team ever built. We duplicated the guide so new users and existing users were asked separately, which kept the two populations clean for analysis. Within a month the product team could say which role submitted the most requests, which role dropped out of the request form, and which role never opened the help content, and target guides at each. Nobody wrote a line of code.

The same team also segmented by how accounts arrived: migrated from an acquired product, self-serve signups, and enterprise contracts. Those three segments explained more of the usage differences than any feature-level report had, because the groups were living in different versions of the product’s history.

Build your first segment with simple filters

Start with one condition, not five. Something like accounts on the Enterprise plan, or visitors who signed up in the last 30 days. Save it. Look at it. Add a second condition only if the first one told you something useful.

This is where people get overwhelmed, trying to build the perfect filter stack before they have looked at a single result. Starting simple only works if you stop and read what the first filter shows you. Skip that step and you are collecting unused segments, the same way every team has a folder called to sort later.

One setting to get right at the start: visibility. A segment set to Only Me cannot be used for guide targeting or nested inside another segment. Set it to Everyone if anyone else will ever need it.

Add metadata for context

Metadata is the extra information attached to each visitor or account that Pendo did not collect on its own: plan type, industry, renewal date, role. It usually comes from your install snippet or from a CRM such as Salesforce or HubSpot. Metadata is what turns these 40 users did X into these 40 users on the Growth plan, past their 90-day mark, did X. The second sentence is the one you can act on. Where each field should come from, and who owns it, is part of the first-90-days plan in the Pendo admin setup post.

Review and refine your segments

Segments go stale. One built around a feature you launched two quarters ago may not matter now, and if a metadata field a segment depends on gets deleted, the segment becomes invalid and any guide using it stops showing. Set a reminder, monthly or quarterly, to look at every saved segment and ask if it still earns its spot. When was the last time you deleted a segment instead of adding another one?

If you are already running NPS, the Pendo NPS post covers how to read scores by segment, which is the natural next step once the segments exist.

The real work is the system around it

Segmentation by itself does not fix anything. A clean segment handed to nobody with a plan to act on it is a well-organized spreadsheet nobody reads. The teams that get value from Pendo data segmentation already decided who owns the data, what good looks like for their product, and what happens after a segment surfaces an answer. That decision layer is what our analytics and reporting work builds, and it is worth having before you go deeper into segments.

A segment is only as useful as the decision it is attached to.

Questions people ask

What is the difference between a segment and metadata in Pendo?

Metadata is the raw information attached to a visitor or account, like plan type or signup date. A segment is the saved filter you build using that metadata and activity data to isolate a group you want to study or target.

How many segments should a beginner build?

Start with two or three. One based on account type and one based on activity level is usually enough to get useful answers before you add more. The role segment from a one-question guide is a strong third.

Can I change a segment after I create it?

Yes. Segments in Pendo are meant to be edited as your questions change. Review them regularly, and remember that deleting a metadata field a segment uses will break that segment and any guide targeted with it.

Build one segment this week, based on a question you already ask out loud in meetings. If you would rather have someone look at your metadata first, book a free Pendo strategy call and we will tell you which three fields to start with.

Put this to work

Want a second opinion on your Pendo setup?

A free 30-minute strategy call. We look at your instance, name the biggest leak, and tell you what we would do first. No deck, no pitch.

Book a free strategy call